BWA - input files

Bruins · 07-06-2010, 08:08 AM

Hi,

I am trying to align paired end reads (Illumina) to a reference genome using BWA. I have 10 reads files, 5 for each direction.

In this old post 'totalnew' says to align the files separately:

Code:

bwa aln database.fasta 4_1.fq > 1_1.fq.sai
bwa aln database.fasta 4_2.fq > 1_2.fq.sai
bwa aln database.fasta 5_1.fq > 2_1.fq.sai
bwa aln database.fasta 5_2.fq > 2_2.fq.sai
[...]

I have been reading for quite a while now so sorry if I missed something totally obvious but how do I run sampe from there on? Can I just say

Code:

bwa sampe database.fa 1_1.fq.sai 2_1.fq.sai 3_1... 1_2.fq.sai 2_2.fq.sai 3_2... 1_1.fq 2_1.fq 3_1... 1_2.fq 2_2.fq 3_2... > alignment.sam

or do I need to run sampe for each file separately?
What if I concatenate all reads files into s_1_sequence.txt and s_2_sequences.txt and then run bwa aln twice and bwa samse once?

cheers!

ps (offtopic) @lh3: I tried downloading bwa 0.5.8a but it seems as though there are files missing. Here what I get:

Code:

wget http://sourceforge.net/projects/bio-bwa/files/bwa-0.5.8a.tar.bz2/download
bunzip2 bwa-0.5.8a.tar.bz2
tar -xf bwa-0.5.8a.tar
ls bwa-0.5.8a
bntseq.c  bwase.c     bwtaln.c  bwtgap.h    bwtio.c     bwtsw2_aux.c    bwtsw2_main.c  is.c     kstring.c  main.h        simple_dp.c     utils.c
bntseq.h  bwase.h     bwtaln.h  bwt_gen     bwt_lite.c  bwtsw2_chain.c  ChangeLog      khash.h  kstring.h  Makefile      solid2fastq.pl  utils.h
bwa.1     bwaseqio.c  bwt.c     bwt.h       bwt_lite.h  bwtsw2_core.c   COPYING        kseq.h   kvec.h     NEWS          stdaln.c
bwape.c   bwa.txt     bwtgap.c  bwtindex.c  bwtmisc.c   bwtsw2.h        cs2nt.c        ksort.h  main.c     qualfa2fq.pl  stdaln.h

I downloaded 0.5.7 and it runs fine.

BENM · 07-06-2010, 08:39 AM

I think you'd better to run sampe for each PE files separately:

bwa sampe [options] <prefix> <in1.sai> <in2.sai> <in1.fq> <in2.fq>

Although you concatenate all read files into two PE seq is okay, I recommend you split them (for example: 100M reads per files) and run in different CPU cores or computer nodes (MPI mode), so that it would be more faster.

Bruins · 07-07-2010, 12:43 AM

Hi,

Thanks for your reply.

So you suggest running sampe at least five times and then merge the resulting sam files?

I will try this.

Cheers!

*** edit
Thanks, works fine.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Several input files w. Novoalign?	kga1978	Bioinformatics	18	11-23-2011 01:57 AM
where is the error in my input files?	shuang	Bioinformatics	3	08-23-2011 02:23 AM
input files for IMAGE	Maegwin	Bioinformatics	4	04-22-2011 05:54 PM
SVA input files	srd	Introductions	0	03-16-2011 07:17 AM
IMAGE input files	skingan	Genomic Resequencing	0	07-29-2010 01:02 PM

07-07-2010, 12:43 AM	#3
Bruins Member Location: Groningen Join Date: Feb 2010 Posts: 78	Hi, Thanks for your reply. So you suggest running sampe at least five times and then merge the resulting sam files? I will try this. Cheers! *** edit Thanks, works fine. Last edited by Bruins; 07-12-2010 at 02:47 AM.

07-06-2010, 08:39 AM	#2
BENM Member Location: PRC Join Date: May 2009 Posts: 33	I think you'd better to run sampe for each PE files separately: bwa sampe [options] <prefix> <in1.sai> <in2.sai> <in1.fq> <in2.fq> Although you concatenate all read files into two PE seq is okay, I recommend you split them (for example: 100M reads per files) and run in different CPU cores or computer nodes (MPI mode), so that it would be more faster.